Hello,

I have a Coda server running on Linux (RedHat 6.2) in a LAN, and also
a few RedHat clients, and everything seems to work very nicely.

However, I also want to access the server from a box outside of the
LAN (across the 'net), which is running Linux 2.3.99-pre3 with the
pre6-7 coda patches applied.

The point is that from this machine, I can access the Coda testserver
without any problems.  But when I access the main server, I cannot
access any files.

Venus starts up and tells me that it has mounted /coda, but the first
access just hangs for 30 seconds and then returns a timeout.  (I can
also authenticate to Coda with clog.)

This is part of codacon output (with the actual machine name replaced
by a fake name...):

connection::bandwidth myserver.domain.se 21712 36813 120889 ( 11:02:16 )
Fetch (0x7f000000.0x1.0x1) [2] ( 11:02:48 )
progress::fetching () 0x ( 11:02:48 )
connection::unreachable myserver.domain.se ( 11:03:08 )


And here is some Venus logging output:

[ W(20) : 0000 : 11:03:08 ] CommQueue: remove pri 3 count = 0
[ W(20) : 0000 : 11:03:08 ] (Multi)Fetch: code = -2001, elapsed = 20000.1
[ W(20) : 0000 : 11:03:08 ] RPC2_SStats: Total = 0
[ W(20) : 0000 : 11:03:08 ] SFTP_SStats: Starts = 0, Datas = 0, DataRetries = 0, A
cks = 0
[ W(20) : 0000 : 11:03:08 ] RPC2_RStats: Replies = 0, Busies = 0, Naks = 0, Bogus= 0
[ W(20) : 0000 : 11:03:08 ] SFTP_RStats: Datas = 0, Acks = 0, Busies = 0
[ W(20) : 0000 : 11:03:08 ] *** Long Running (Multi)Fetch: code = -2001, elapsed =
 20000.1 ***
[ W(20) : 0000 : 11:03:08 ] mgrpent::CheckNonMutating: acode = -2001
                hosts = [0xc0a80102 0 0 0 0 0 0 0],
                retcodes = [-2015 -2002 -2002 -2002 -2002 -2002 -2002 -2002]
[ W(20) : 0000 : 11:03:08 ] GetServer: host = c0a80102
[ W(20) : 0000 : 11:03:08 ] srvent::ServerError: myserver.domain.se error (RPC2_DEAD (F))
[ W(20) : 0000 : 11:03:08 ] srvent::Reset: host = myserver.domain.se
[ W(20) : 0000 : 11:03:08 ] connent::Suicide: disconnect = 0


Note that it seems to work perfectly with the testserver.  In that
case the first (Multi)Fetch line gives the same code (-2001) but a
shorter time interval.  Network connectivity is a lot better to my
server than the testserver, so that should not be the problem.

So I cannot even figure out if the problem is on the client or the
server... Any help would be appreciated.

/Marcus B.